Date of Award

Spring 1-1-2018

Document Type


Degree Name

Doctor of Philosophy (PhD)

First Advisor

Christine Hrenya

Second Advisor

Xiaolong Yin

Third Advisor

Michael Shirts

Fourth Advisor

Alan Weimer

Fifth Advisor

Richard Regueiro


Despite the remarkable advancements made by modern science, multi-scale problems still pose a significant challenge to the fields of engineering and mathematics. The transport phenomena occurring within a gas-solids flow is a quintessential multi-scale problem. Specifically, the physics occurring at the sub-particle scale is strongly coupled to the macroscopic behavior of many flowing particles. In addition to a fundamental significance, the mathematical desciption of a particle-laden flow is crucial to the efficient design and operation of many industrial systems. Products from multiphase unit operations encompass (but are not limited to): petroleum, pharmaceuticals, polymers, limestone, and energy. In recent years, solid particles have played a key role in the area of renewable energy. Namely, efficient designs for concentrated solar power plants (CSPs) have been proposed that utilize solid particles as the heat transfer fluid. Particle-based CSPs employ a near black body (NBB) receiver (i.e., granular heat exchanger) to transfer thermal energy from concentrated sunlight to a gas-solids mixture. Generally, concentrated sunlight is irradiated upon the NBB domain walls and the solid particles undergo heat transfer with the hot walls as they flow through the receiver. To assess the feasibility of these new CSP designs, the accurate prediction of wall-to-particle heat transfer is of primary significance. Due to the novelty of solid particle CSPs, the first steps in quantifying wall-to-particle heat transfer have been concerned with convective and conductive mechanisms only (no radiation). As a result, the present work will be concerned with the convective and conductive transport of thermal energy within a wall-bounded gas-solids flow.

The work here begins with an application-based study of the heat transfer within a NBB receiver. The discrete element method (DEM) is employed to simulate a granular heat exchanger whose domain walls are exposed to a constant total heat flux (irradiated sunlight). Current state-of-the-art methods for simulating wall-to-particle heat transfer were implemented within the DEM framework. Specifically, closures for the direct conduction mechanism occurring between a particle and wall in contact and indirect conduction mechanism occurring between a particle and wall separated by a thin layer of fluid were added to DEM. Furthermore, a new, total heat flux boundary condition was developed to properly describe the NBB geometry. Previous boundary conditions required that a heat flux be specified to each phase and leads to different wall temperatures for each phase. Physically speaking, the total heat flux at the NBB domain wall is what may be approximated (irradiated sunlight) while the partition of the total heat flux amongst the gas and solids phase will vary in space and time according to the local hydrodynamics. A numerical framework for implementing the latter interpretation for a constant, total heat flux boundary condition is derived and verified against DEM simulation of four granular heat exchanger designs.

Simulation of the NBB receiver showed that a majority of the heat transfer to the particles is due to the indirect conduction mechanism. A comparison of the thermal resistances associated with direct and indirect conduction is formulated and it is noted that indirect conduction will dominate for a wide variety of systems. In light of the significance of indirect conduction, the sensitivity of indirect conduction theory to its two theoretical inputs (fluid lens thickness and surface roughness) is assessed for dynamic, multi-particle systems. Analytical techniques commonly employed by kinetic theory are utilized to average indirect conduction theory and quantify a macroscopic heat transfer coefficient. Inputs to indirect conduction theory (fluid lens thickness and surface roughness) are perturbed individually to quantify their effect upon th